The 3D semantic labeling task involves predicting a semantic labeling of a 3D scan mesh.

Evaluation and metrics

Our evaluation ranks all methods according to the PASCAL VOC intersection-over-union metric (IoU). IoU = TP/(TP+FP+FN), where TP, FP, and FN are the numbers of true positive, false positive, and false negative pixels, respectively. Predicted labels are evaluated per-vertex over the respective 3D scan mesh; for 3D approaches that operate on other representations like grids or points, the predicted labels should be mapped onto the mesh vertices (e.g., one such example for grid to mesh vertices is provided in the evaluation helpers).



This table lists the benchmark results for the 3D semantic label scenario.


Method Infoavg ioubathtubbedbookshelfcabinetchaircountercurtaindeskdoorfloorotherfurniturepicturerefrigeratorshower curtainsinksofatabletoiletwallwindow
sorted bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort by
PTv3-PPT-ALCcopyleft0.798 10.911 100.812 210.854 70.770 120.856 140.555 150.943 10.660 240.735 20.979 10.606 70.492 10.792 40.934 30.841 20.819 50.716 80.947 100.906 10.822 1
PTv3 ScanNet0.794 20.941 30.813 200.851 90.782 60.890 30.597 10.916 50.696 90.713 50.979 10.635 20.384 30.793 30.907 100.821 50.790 330.696 130.967 30.903 20.805 2
Xiaoyang Wu, Li Jiang, Peng-Shuai Wang, Zhijian Liu, Xihui Liu, Yu Qiao, Wanli Ouyang, Tong He, Hengshuang Zhao: Point Transformer V3: Simpler, Faster, Stronger. CVPR 2024 (Oral)
DITR ScanNet0.793 30.811 400.852 20.889 10.774 90.907 10.592 20.927 30.719 10.718 30.961 170.652 10.348 120.817 10.927 50.795 90.824 20.749 10.948 90.887 70.771 11
PonderV20.785 40.978 10.800 290.833 260.788 40.853 190.545 190.910 80.713 20.705 60.979 10.596 90.390 20.769 150.832 440.821 50.792 320.730 20.975 10.897 50.785 6
Haoyi Zhu, Honghui Yang, Xiaoyang Wu, Di Huang, Sha Zhang, Xianglong He, Tong He, Hengshuang Zhao, Chunhua Shen, Yu Qiao, Wanli Ouyang: PonderV2: Pave the Way for 3D Foundataion Model with A Universal Pre-training Paradigm.
Mix3Dpermissive0.781 50.964 20.855 10.843 180.781 70.858 130.575 70.831 360.685 150.714 40.979 10.594 100.310 290.801 20.892 180.841 20.819 50.723 50.940 150.887 70.725 27
Alexey Nekrasov, Jonas Schult, Or Litany, Bastian Leibe, Francis Engelmann: Mix3D: Out-of-Context Data Augmentation for 3D Scenes. 3DV 2021 (Oral)
Swin3Dpermissive0.779 60.861 220.818 150.836 230.790 30.875 50.576 60.905 90.704 60.739 10.969 110.611 30.349 110.756 250.958 10.702 480.805 160.708 90.916 350.898 40.801 3
TTT-KD0.773 70.646 940.818 150.809 380.774 90.878 40.581 30.943 10.687 130.704 70.978 50.607 60.336 180.775 110.912 80.838 40.823 30.694 140.967 30.899 30.794 5
Lisa Weijler, Muhammad Jehanzeb Mirza, Leon Sick, Can Ekkazan, Pedro Hermosilla: TTT-KD: Test-Time Training for 3D Semantic Segmentation through Knowledge Distillation from Foundation Models.
ResLFE_HDS0.772 80.939 40.824 70.854 70.771 110.840 330.564 110.900 110.686 140.677 140.961 170.537 340.348 120.769 150.903 120.785 130.815 80.676 250.939 160.880 130.772 10
OctFormerpermissive0.766 90.925 70.808 250.849 110.786 50.846 290.566 100.876 180.690 110.674 160.960 190.576 200.226 700.753 270.904 110.777 150.815 80.722 60.923 300.877 160.776 9
Peng-Shuai Wang: OctFormer: Octree-based Transformers for 3D Point Clouds. SIGGRAPH 2023
PPT-SpUNet-Joint0.766 90.932 50.794 350.829 280.751 250.854 170.540 230.903 100.630 370.672 170.963 150.565 240.357 90.788 50.900 140.737 280.802 170.685 190.950 70.887 70.780 7
Xiaoyang Wu, Zhuotao Tian, Xin Wen, Bohao Peng, Xihui Liu, Kaicheng Yu, Hengshuang Zhao: Towards Large-scale 3D Representation Learning with Multi-dataset Point Prompt Training. CVPR 2024
OccuSeg+Semantic0.764 110.758 610.796 330.839 210.746 280.907 10.562 120.850 280.680 170.672 170.978 50.610 40.335 200.777 90.819 480.847 10.830 10.691 160.972 20.885 100.727 25
CU-Hybrid Net0.764 110.924 80.819 130.840 200.757 200.853 190.580 40.848 290.709 40.643 270.958 230.587 150.295 360.753 270.884 220.758 220.815 80.725 40.927 260.867 250.743 18
O-CNNpermissive0.762 130.924 80.823 80.844 170.770 120.852 210.577 50.847 310.711 30.640 310.958 230.592 110.217 760.762 200.888 190.758 220.813 120.726 30.932 240.868 240.744 17
Peng-Shuai Wang, Yang Liu, Yu-Xiao Guo, Chun-Yu Sun, Xin Tong: O-CNN: Octree-based Convolutional Neural Networks for 3D Shape Analysis. SIGGRAPH 2017
DiffSegNet0.758 140.725 770.789 400.843 180.762 160.856 140.562 120.920 40.657 270.658 210.958 230.589 130.337 170.782 60.879 230.787 110.779 380.678 210.926 280.880 130.799 4
DTC0.757 150.843 280.820 110.847 140.791 20.862 110.511 360.870 200.707 50.652 230.954 380.604 80.279 470.760 210.942 20.734 290.766 470.701 120.884 570.874 220.736 19
OA-CNN-L_ScanNet200.756 160.783 470.826 60.858 50.776 80.837 360.548 180.896 140.649 290.675 150.962 160.586 160.335 200.771 140.802 520.770 180.787 350.691 160.936 190.880 130.761 13
PNE0.755 170.786 450.835 50.834 250.758 180.849 240.570 90.836 350.648 300.668 190.978 50.581 190.367 70.683 380.856 320.804 70.801 210.678 210.961 50.889 60.716 32
P. Hermosilla: Point Neighborhood Embeddings.
ConDaFormer0.755 170.927 60.822 90.836 230.801 10.849 240.516 330.864 250.651 280.680 130.958 230.584 180.282 440.759 230.855 340.728 310.802 170.678 210.880 620.873 230.756 15
Lunhao Duan, Shanshan Zhao, Nan Xue, Mingming Gong, Guisong Xia, Dacheng Tao: ConDaFormer : Disassembled Transformer with Local Structure Enhancement for 3D Point Cloud Understanding. Neurips, 2023
PointTransformerV20.752 190.742 680.809 240.872 20.758 180.860 120.552 160.891 160.610 440.687 80.960 190.559 280.304 320.766 180.926 60.767 190.797 250.644 360.942 130.876 190.722 29
Xiaoyang Wu, Yixing Lao, Li Jiang, Xihui Liu, Hengshuang Zhao: Point Transformer V2: Grouped Vector Attention and Partition-based Pooling. NeurIPS 2022
DMF-Net0.752 190.906 130.793 370.802 440.689 430.825 490.556 140.867 210.681 160.602 470.960 190.555 300.365 80.779 80.859 290.747 250.795 290.717 70.917 340.856 330.764 12
C.Yang, Y.Yan, W.Zhao, J.Ye, X.Yang, A.Hussain, B.Dong, K.Huang: Towards Deeper and Better Multi-view Feature Fusion for 3D Semantic Segmentation. ICONIP 2023
PointConvFormer0.749 210.793 430.790 380.807 400.750 270.856 140.524 290.881 170.588 560.642 300.977 90.591 120.274 500.781 70.929 40.804 70.796 260.642 370.947 100.885 100.715 33
Wenxuan Wu, Qi Shan, Li Fuxin: PointConvFormer: Revenge of the Point-based Convolution.
BPNetcopyleft0.749 210.909 110.818 150.811 360.752 230.839 350.485 500.842 320.673 190.644 260.957 280.528 400.305 310.773 120.859 290.788 100.818 70.693 150.916 350.856 330.723 28
Wenbo Hu, Hengshuang Zhao, Li Jiang, Jiaya Jia, Tien-Tsin Wong: Bidirectional Projection Network for Cross Dimension Scene Understanding. CVPR 2021 (Oral)
MSP0.748 230.623 970.804 270.859 40.745 290.824 510.501 400.912 70.690 110.685 100.956 290.567 230.320 260.768 170.918 70.720 360.802 170.676 250.921 320.881 120.779 8
StratifiedFormerpermissive0.747 240.901 140.803 280.845 160.757 200.846 290.512 350.825 390.696 90.645 250.956 290.576 200.262 610.744 320.861 280.742 260.770 450.705 100.899 470.860 300.734 20
Xin Lai*, Jianhui Liu*, Li Jiang, Liwei Wang, Hengshuang Zhao, Shu Liu, Xiaojuan Qi, Jiaya Jia: Stratified Transformer for 3D Point Cloud Segmentation. CVPR 2022
Virtual MVFusion0.746 250.771 550.819 130.848 130.702 400.865 100.397 880.899 120.699 70.664 200.948 580.588 140.330 220.746 310.851 380.764 200.796 260.704 110.935 200.866 260.728 23
Abhijit Kundu, Xiaoqi Yin, Alireza Fathi, David Ross, Brian Brewington, Thomas Funkhouser, Caroline Pantofaru: Virtual Multi-view Fusion for 3D Semantic Segmentation. ECCV 2020
VMNetpermissive0.746 250.870 200.838 30.858 50.729 340.850 230.501 400.874 190.587 570.658 210.956 290.564 250.299 340.765 190.900 140.716 390.812 130.631 420.939 160.858 310.709 34
Zeyu HU, Xuyang Bai, Jiaxiang Shang, Runze Zhang, Jiayu Dong, Xin Wang, Guangyuan Sun, Hongbo Fu, Chiew-Lan Tai: VMNet: Voxel-Mesh Network for Geodesic-Aware 3D Semantic Segmentation. ICCV 2021 (Oral)
DiffSeg3D20.745 270.725 770.814 190.837 220.751 250.831 430.514 340.896 140.674 180.684 110.960 190.564 250.303 330.773 120.820 470.713 420.798 240.690 180.923 300.875 200.757 14
Retro-FPN0.744 280.842 290.800 290.767 580.740 300.836 380.541 210.914 60.672 200.626 350.958 230.552 310.272 520.777 90.886 210.696 490.801 210.674 280.941 140.858 310.717 30
Peng Xiang*, Xin Wen*, Yu-Shen Liu, Hui Zhang, Yi Fang, Zhizhong Han: Retrospective Feature Pyramid Network for Point Cloud Semantic Segmentation. ICCV 2023
EQ-Net0.743 290.620 980.799 320.849 110.730 330.822 530.493 470.897 130.664 210.681 120.955 320.562 270.378 40.760 210.903 120.738 270.801 210.673 290.907 390.877 160.745 16
Zetong Yang*, Li Jiang*, Yanan Sun, Bernt Schiele, Jiaya JIa: A Unified Query-based Paradigm for Point Cloud Understanding. CVPR 2022
LRPNet0.742 300.816 370.806 260.807 400.752 230.828 470.575 70.839 340.699 70.637 320.954 380.520 430.320 260.755 260.834 420.760 210.772 420.676 250.915 370.862 280.717 30
SAT0.742 300.860 230.765 520.819 310.769 140.848 260.533 250.829 370.663 220.631 340.955 320.586 160.274 500.753 270.896 160.729 300.760 530.666 310.921 320.855 350.733 21
LargeKernel3D0.739 320.909 110.820 110.806 420.740 300.852 210.545 190.826 380.594 550.643 270.955 320.541 330.263 600.723 360.858 310.775 170.767 460.678 210.933 220.848 400.694 39
Yukang Chen*, Jianhui Liu*, Xiangyu Zhang, Xiaojuan Qi, Jiaya Jia: LargeKernel3D: Scaling up Kernels in 3D Sparse CNNs. CVPR 2023
RPN0.736 330.776 510.790 380.851 90.754 220.854 170.491 490.866 230.596 540.686 90.955 320.536 350.342 150.624 530.869 250.787 110.802 170.628 430.927 260.875 200.704 36
MinkowskiNetpermissive0.736 330.859 240.818 150.832 270.709 380.840 330.521 310.853 270.660 240.643 270.951 480.544 320.286 420.731 340.893 170.675 580.772 420.683 200.874 690.852 380.727 25
C. Choy, J. Gwak, S. Savarese: 4D Spatio-Temporal ConvNets: Minkowski Convolutional Neural Networks. CVPR 2019
IPCA0.731 350.890 160.837 40.864 30.726 350.873 60.530 280.824 400.489 900.647 240.978 50.609 50.336 180.624 530.733 610.758 220.776 400.570 680.949 80.877 160.728 23
SparseConvNet0.725 360.647 930.821 100.846 150.721 360.869 70.533 250.754 610.603 500.614 390.955 320.572 220.325 240.710 370.870 240.724 340.823 30.628 430.934 210.865 270.683 42
PointTransformer++0.725 360.727 760.811 230.819 310.765 150.841 320.502 390.814 450.621 400.623 370.955 320.556 290.284 430.620 550.866 260.781 140.757 570.648 340.932 240.862 280.709 34
MatchingNet0.724 380.812 390.812 210.810 370.735 320.834 400.495 460.860 260.572 640.602 470.954 380.512 450.280 460.757 240.845 400.725 330.780 370.606 530.937 180.851 390.700 38
INS-Conv-semantic0.717 390.751 640.759 560.812 350.704 390.868 80.537 240.842 320.609 460.608 430.953 420.534 370.293 370.616 560.864 270.719 380.793 300.640 380.933 220.845 440.663 48
PointMetaBase0.714 400.835 300.785 410.821 290.684 450.846 290.531 270.865 240.614 410.596 510.953 420.500 480.246 660.674 390.888 190.692 500.764 490.624 450.849 850.844 450.675 44
contrastBoundarypermissive0.705 410.769 580.775 460.809 380.687 440.820 560.439 760.812 460.661 230.591 530.945 660.515 440.171 950.633 500.856 320.720 360.796 260.668 300.889 540.847 410.689 40
Liyao Tang, Yibing Zhan, Zhe Chen, Baosheng Yu, Dacheng Tao: Contrastive Boundary Learning for Point Cloud Segmentation. CVPR2022
ClickSeg_Semantic0.703 420.774 530.800 290.793 490.760 170.847 280.471 540.802 490.463 970.634 330.968 130.491 510.271 540.726 350.910 90.706 440.815 80.551 800.878 630.833 460.570 80
RFCR0.702 430.889 170.745 670.813 340.672 480.818 610.493 470.815 440.623 380.610 410.947 600.470 600.249 650.594 590.848 390.705 450.779 380.646 350.892 520.823 520.611 63
Jingyu Gong, Jiachen Xu, Xin Tan, Haichuan Song, Yanyun Qu, Yuan Xie, Lizhuang Ma: Omni-Supervised Point Cloud Segmentation via Gradual Receptive Field Component Reasoning. CVPR2021
One Thing One Click0.701 440.825 340.796 330.723 650.716 370.832 420.433 780.816 420.634 350.609 420.969 110.418 860.344 140.559 710.833 430.715 400.808 150.560 740.902 440.847 410.680 43
JSENetpermissive0.699 450.881 190.762 530.821 290.667 490.800 730.522 300.792 520.613 420.607 440.935 860.492 500.205 810.576 640.853 360.691 520.758 550.652 330.872 720.828 490.649 52
Zeyu HU, Mingmin Zhen, Xuyang BAI, Hongbo Fu, Chiew-lan Tai: JSENet: Joint Semantic Segmentation and Edge Detection Network for 3D Point Clouds. ECCV 2020
One-Thing-One-Click0.693 460.743 670.794 350.655 880.684 450.822 530.497 450.719 710.622 390.617 380.977 90.447 730.339 160.750 300.664 780.703 470.790 330.596 580.946 120.855 350.647 53
Zhengzhe Liu, Xiaojuan Qi, Chi-Wing Fu: One Thing One Click: A Self-Training Approach for Weakly Supervised 3D Semantic Segmentation. CVPR 2021
PicassoNet-IIpermissive0.692 470.732 720.772 470.786 500.677 470.866 90.517 320.848 290.509 830.626 350.952 460.536 350.225 720.545 770.704 690.689 550.810 140.564 730.903 430.854 370.729 22
Huan Lei, Naveed Akhtar, Mubarak Shah, and Ajmal Mian: Geometric feature learning for 3D meshes.
Feature_GeometricNetpermissive0.690 480.884 180.754 600.795 470.647 560.818 610.422 800.802 490.612 430.604 450.945 660.462 630.189 890.563 700.853 360.726 320.765 480.632 410.904 410.821 550.606 67
Kangcheng Liu, Ben M. Chen: https://arxiv.org/abs/2012.09439. arXiv Preprint
FusionNet0.688 490.704 830.741 710.754 620.656 510.829 450.501 400.741 660.609 460.548 610.950 520.522 420.371 50.633 500.756 560.715 400.771 440.623 460.861 800.814 580.658 49
Feihu Zhang, Jin Fang, Benjamin Wah, Philip Torr: Deep FusionNet for Point Cloud Semantic Segmentation. ECCV 2020
Feature-Geometry Netpermissive0.685 500.866 210.748 640.819 310.645 580.794 760.450 660.802 490.587 570.604 450.945 660.464 620.201 840.554 730.840 410.723 350.732 680.602 560.907 390.822 540.603 70
DGNet0.684 510.712 820.784 420.782 540.658 500.835 390.499 440.823 410.641 320.597 500.950 520.487 530.281 450.575 650.619 820.647 710.764 490.620 480.871 750.846 430.688 41
VACNN++0.684 510.728 750.757 590.776 550.690 410.804 710.464 590.816 420.577 630.587 540.945 660.508 470.276 490.671 400.710 670.663 630.750 610.589 630.881 600.832 480.653 51
KP-FCNN0.684 510.847 270.758 580.784 520.647 560.814 640.473 530.772 550.605 480.594 520.935 860.450 710.181 920.587 600.805 510.690 530.785 360.614 490.882 590.819 560.632 59
H. Thomas, C. Qi, J. Deschaud, B. Marcotegui, F. Goulette, L. Guibas.: KPConv: Flexible and Deformable Convolution for Point Clouds. ICCV 2019
PointContrast_LA_SEM0.683 540.757 620.784 420.786 500.639 600.824 510.408 830.775 540.604 490.541 630.934 900.532 380.269 560.552 740.777 540.645 740.793 300.640 380.913 380.824 510.671 45
Superpoint Network0.683 540.851 260.728 750.800 460.653 530.806 690.468 560.804 470.572 640.602 470.946 630.453 700.239 690.519 830.822 450.689 550.762 520.595 600.895 500.827 500.630 60
VI-PointConv0.676 560.770 570.754 600.783 530.621 640.814 640.552 160.758 590.571 660.557 590.954 380.529 390.268 580.530 800.682 730.675 580.719 710.603 550.888 550.833 460.665 47
Xingyi Li, Wenxuan Wu, Xiaoli Z. Fern, Li Fuxin: The Devils in the Point Clouds: Studying the Robustness of Point Cloud Convolutions.
ROSMRF3D0.673 570.789 440.748 640.763 600.635 620.814 640.407 850.747 630.581 610.573 560.950 520.484 540.271 540.607 570.754 570.649 680.774 410.596 580.883 580.823 520.606 67
SALANet0.670 580.816 370.770 500.768 570.652 540.807 680.451 630.747 630.659 260.545 620.924 970.473 590.149 1050.571 670.811 500.635 770.746 620.623 460.892 520.794 710.570 80
O3DSeg0.668 590.822 350.771 490.496 1090.651 550.833 410.541 210.761 580.555 720.611 400.966 140.489 520.370 60.388 1020.580 850.776 160.751 590.570 680.956 60.817 570.646 54
PointConvpermissive0.666 600.781 480.759 560.699 730.644 590.822 530.475 520.779 530.564 690.504 800.953 420.428 800.203 830.586 620.754 570.661 640.753 580.588 640.902 440.813 600.642 55
Wenxuan Wu, Zhongang Qi, Li Fuxin: PointConv: Deep Convolutional Networks on 3D Point Clouds. CVPR 2019
PointASNLpermissive0.666 600.703 840.781 440.751 640.655 520.830 440.471 540.769 560.474 930.537 650.951 480.475 580.279 470.635 480.698 720.675 580.751 590.553 790.816 920.806 620.703 37
Xu Yan, Chaoda Zheng, Zhen Li, Sheng Wang, Shuguang Cui: PointASNL: Robust Point Clouds Processing using Nonlocal Neural Networks with Adaptive Sampling. CVPR 2020
PPCNN++permissive0.663 620.746 650.708 780.722 660.638 610.820 560.451 630.566 990.599 520.541 630.950 520.510 460.313 280.648 450.819 480.616 820.682 860.590 620.869 760.810 610.656 50
Pyunghwan Ahn, Juyoung Yang, Eojindl Yi, Chanho Lee, Junmo Kim: Projection-based Point Convolution for Efficient Point Cloud Segmentation. IEEE Access
DCM-Net0.658 630.778 490.702 810.806 420.619 650.813 670.468 560.693 790.494 860.524 720.941 780.449 720.298 350.510 850.821 460.675 580.727 700.568 710.826 900.803 650.637 57
Jonas Schult*, Francis Engelmann*, Theodora Kontogianni, Bastian Leibe: DualConvMesh-Net: Joint Geodesic and Euclidean Convolutions on 3D Meshes. CVPR 2020 [Oral]
MVF-GNN0.658 630.558 1050.751 620.655 880.690 410.722 980.453 620.867 210.579 620.576 550.893 1090.523 410.293 370.733 330.571 870.692 500.659 930.606 530.875 660.804 640.668 46
HPGCNN0.656 650.698 860.743 690.650 900.564 820.820 560.505 380.758 590.631 360.479 840.945 660.480 560.226 700.572 660.774 550.690 530.735 660.614 490.853 840.776 860.597 73
Jisheng Dang, Qingyong Hu, Yulan Guo, Jun Yang: HPGCNN.
SAFNet-segpermissive0.654 660.752 630.734 730.664 860.583 770.815 630.399 870.754 610.639 330.535 670.942 760.470 600.309 300.665 410.539 890.650 670.708 760.635 400.857 830.793 730.642 55
Linqing Zhao, Jiwen Lu, Jie Zhou: Similarity-Aware Fusion Network for 3D Semantic Segmentation. IROS 2021
RandLA-Netpermissive0.645 670.778 490.731 740.699 730.577 780.829 450.446 680.736 670.477 920.523 740.945 660.454 670.269 560.484 920.749 600.618 800.738 630.599 570.827 890.792 760.621 62
PointConv-SFPN0.641 680.776 510.703 800.721 670.557 850.826 480.451 630.672 850.563 700.483 830.943 750.425 830.162 1000.644 460.726 620.659 650.709 750.572 670.875 660.786 810.559 86
MVPNetpermissive0.641 680.831 310.715 760.671 830.590 730.781 820.394 890.679 820.642 310.553 600.937 830.462 630.256 620.649 440.406 1020.626 780.691 830.666 310.877 640.792 760.608 66
Maximilian Jaritz, Jiayuan Gu, Hao Su: Multi-view PointNet for 3D Scene Understanding. GMDL Workshop, ICCV 2019
PointMRNet0.640 700.717 810.701 820.692 760.576 790.801 720.467 580.716 720.563 700.459 900.953 420.429 790.169 970.581 630.854 350.605 830.710 730.550 810.894 510.793 730.575 78
FPConvpermissive0.639 710.785 460.760 550.713 710.603 680.798 740.392 900.534 1040.603 500.524 720.948 580.457 650.250 640.538 780.723 650.598 870.696 810.614 490.872 720.799 660.567 83
Yiqun Lin, Zizheng Yan, Haibin Huang, Dong Du, Ligang Liu, Shuguang Cui, Xiaoguang Han: FPConv: Learning Local Flattening for Point Convolution. CVPR 2020
PD-Net0.638 720.797 420.769 510.641 950.590 730.820 560.461 600.537 1030.637 340.536 660.947 600.388 930.206 800.656 420.668 760.647 710.732 680.585 650.868 770.793 730.473 106
PointSPNet0.637 730.734 710.692 890.714 700.576 790.797 750.446 680.743 650.598 530.437 950.942 760.403 890.150 1040.626 520.800 530.649 680.697 800.557 770.846 860.777 850.563 84
SConv0.636 740.830 320.697 850.752 630.572 810.780 840.445 700.716 720.529 760.530 680.951 480.446 740.170 960.507 870.666 770.636 760.682 860.541 870.886 560.799 660.594 74
Supervoxel-CNN0.635 750.656 910.711 770.719 680.613 660.757 930.444 730.765 570.534 750.566 570.928 950.478 570.272 520.636 470.531 910.664 620.645 970.508 950.864 790.792 760.611 63
joint point-basedpermissive0.634 760.614 990.778 450.667 850.633 630.825 490.420 810.804 470.467 950.561 580.951 480.494 490.291 390.566 680.458 970.579 940.764 490.559 760.838 870.814 580.598 72
Hung-Yueh Chiang, Yen-Liang Lin, Yueh-Cheng Liu, Winston H. Hsu: A Unified Point-Based Framework for 3D Segmentation. 3DV 2019
PointMTL0.632 770.731 730.688 920.675 800.591 720.784 810.444 730.565 1000.610 440.492 810.949 560.456 660.254 630.587 600.706 680.599 860.665 920.612 520.868 770.791 790.579 77
PointNet2-SFPN0.631 780.771 550.692 890.672 810.524 900.837 360.440 750.706 770.538 740.446 920.944 720.421 850.219 750.552 740.751 590.591 900.737 640.543 860.901 460.768 890.557 87
APCF-Net0.631 780.742 680.687 940.672 810.557 850.792 790.408 830.665 860.545 730.508 770.952 460.428 800.186 900.634 490.702 700.620 790.706 770.555 780.873 700.798 680.581 76
Haojia, Lin: Adaptive Pyramid Context Fusion for Point Cloud Perception. GRSL
3DSM_DMMF0.631 780.626 960.745 670.801 450.607 670.751 940.506 370.729 700.565 680.491 820.866 1120.434 750.197 870.595 580.630 810.709 430.705 780.560 740.875 660.740 970.491 101
FusionAwareConv0.630 810.604 1010.741 710.766 590.590 730.747 950.501 400.734 680.503 850.527 700.919 1010.454 670.323 250.550 760.420 1010.678 570.688 840.544 840.896 490.795 700.627 61
Jiazhao Zhang, Chenyang Zhu, Lintao Zheng, Kai Xu: Fusion-Aware Point Convolution for Online Semantic 3D Scene Segmentation. CVPR 2020
DenSeR0.628 820.800 410.625 1040.719 680.545 870.806 690.445 700.597 940.448 1000.519 750.938 820.481 550.328 230.489 910.499 960.657 660.759 540.592 610.881 600.797 690.634 58
SegGroup_sempermissive0.627 830.818 360.747 660.701 720.602 690.764 900.385 940.629 910.490 880.508 770.931 940.409 880.201 840.564 690.725 630.618 800.692 820.539 880.873 700.794 710.548 90
An Tao, Yueqi Duan, Yi Wei, Jiwen Lu, Jie Zhou: SegGroup: Seg-Level Supervision for 3D Instance and Semantic Segmentation. TIP 2022
SIConv0.625 840.830 320.694 870.757 610.563 830.772 880.448 670.647 890.520 790.509 760.949 560.431 780.191 880.496 890.614 830.647 710.672 900.535 900.876 650.783 820.571 79
dtc_net0.625 840.703 840.751 620.794 480.535 880.848 260.480 510.676 840.528 770.469 870.944 720.454 670.004 1170.464 940.636 800.704 460.758 550.548 830.924 290.787 800.492 100
HPEIN0.618 860.729 740.668 950.647 920.597 710.766 890.414 820.680 810.520 790.525 710.946 630.432 760.215 770.493 900.599 840.638 750.617 1020.570 680.897 480.806 620.605 69
Li Jiang, Hengshuang Zhao, Shu Liu, Xiaoyong Shen, Chi-Wing Fu, Jiaya Jia: Hierarchical Point-Edge Interaction Network for Point Cloud Semantic Segmentation. ICCV 2019
SPH3D-GCNpermissive0.610 870.858 250.772 470.489 1100.532 890.792 790.404 860.643 900.570 670.507 790.935 860.414 870.046 1140.510 850.702 700.602 850.705 780.549 820.859 810.773 870.534 93
Huan Lei, Naveed Akhtar, and Ajmal Mian: Spherical Kernel for Efficient Graph Convolution on 3D Point Clouds. TPAMI 2020
AttAN0.609 880.760 600.667 960.649 910.521 910.793 770.457 610.648 880.528 770.434 970.947 600.401 900.153 1030.454 950.721 660.648 700.717 720.536 890.904 410.765 900.485 102
Gege Zhang, Qinghua Ma, Licheng Jiao, Fang Liu and Qigong Sun: AttAN: Attention Adversarial Networks for 3D Point Cloud Semantic Segmentation. IJCAI2020
Weakly-Openseg v30.604 890.901 140.762 530.627 970.478 970.820 560.346 1000.689 800.353 1100.528 690.933 910.217 1150.172 940.530 800.725 630.593 890.737 640.515 920.858 820.772 880.515 96
wsss-transformer0.600 900.634 950.743 690.697 750.601 700.781 820.437 770.585 970.493 870.446 920.933 910.394 910.011 1160.654 430.661 790.603 840.733 670.526 910.832 880.761 920.480 103
LAP-D0.594 910.720 790.692 890.637 960.456 1010.773 870.391 920.730 690.587 570.445 940.940 800.381 940.288 400.434 980.453 990.591 900.649 950.581 660.777 960.749 960.610 65
DPC0.592 920.720 790.700 830.602 1010.480 960.762 920.380 950.713 750.585 600.437 950.940 800.369 960.288 400.434 980.509 950.590 920.639 1000.567 720.772 970.755 940.592 75
Francis Engelmann, Theodora Kontogianni, Bastian Leibe: Dilated Point Convolutions: On the Receptive Field Size of Point Convolutions on 3D Point Clouds. ICRA 2020
CCRFNet0.589 930.766 590.659 990.683 780.470 1000.740 970.387 930.620 930.490 880.476 850.922 990.355 990.245 670.511 840.511 940.571 950.643 980.493 990.872 720.762 910.600 71
ROSMRF0.580 940.772 540.707 790.681 790.563 830.764 900.362 970.515 1050.465 960.465 890.936 850.427 820.207 790.438 960.577 860.536 980.675 890.486 1000.723 1030.779 830.524 95
SD-DETR0.576 950.746 650.609 1080.445 1140.517 920.643 1090.366 960.714 740.456 980.468 880.870 1110.432 760.264 590.558 720.674 740.586 930.688 840.482 1010.739 1010.733 990.537 92
SQN_0.1%0.569 960.676 880.696 860.657 870.497 930.779 850.424 790.548 1010.515 810.376 1020.902 1080.422 840.357 90.379 1030.456 980.596 880.659 930.544 840.685 1060.665 1100.556 88
TextureNetpermissive0.566 970.672 900.664 970.671 830.494 940.719 990.445 700.678 830.411 1060.396 1000.935 860.356 980.225 720.412 1000.535 900.565 960.636 1010.464 1030.794 950.680 1070.568 82
Jingwei Huang, Haotian Zhang, Li Yi, Thomas Funkerhouser, Matthias Niessner, Leonidas Guibas: TextureNet: Consistent Local Parametrizations for Learning from High-Resolution Signals on Meshes. CVPR
DVVNet0.562 980.648 920.700 830.770 560.586 760.687 1030.333 1020.650 870.514 820.475 860.906 1050.359 970.223 740.340 1050.442 1000.422 1090.668 910.501 960.708 1040.779 830.534 93
Pointnet++ & Featurepermissive0.557 990.735 700.661 980.686 770.491 950.744 960.392 900.539 1020.451 990.375 1030.946 630.376 950.205 810.403 1010.356 1050.553 970.643 980.497 970.824 910.756 930.515 96
GMLPs0.538 1000.495 1100.693 880.647 920.471 990.793 770.300 1050.477 1060.505 840.358 1040.903 1070.327 1020.081 1110.472 930.529 920.448 1070.710 730.509 930.746 990.737 980.554 89
PanopticFusion-label0.529 1010.491 1110.688 920.604 1000.386 1060.632 1100.225 1160.705 780.434 1030.293 1100.815 1140.348 1000.241 680.499 880.669 750.507 1000.649 950.442 1090.796 940.602 1140.561 85
Gaku Narita, Takashi Seno, Tomoya Ishikawa, Yohsuke Kaji: PanopticFusion: Online Volumetric Semantic Mapping at the Level of Stuff and Things. IROS 2019 (to appear)
subcloud_weak0.516 1020.676 880.591 1110.609 980.442 1020.774 860.335 1010.597 940.422 1050.357 1050.932 930.341 1010.094 1100.298 1070.528 930.473 1050.676 880.495 980.602 1120.721 1020.349 114
Online SegFusion0.515 1030.607 1000.644 1020.579 1030.434 1030.630 1110.353 980.628 920.440 1010.410 980.762 1170.307 1040.167 980.520 820.403 1030.516 990.565 1050.447 1070.678 1070.701 1040.514 98
Davide Menini, Suryansh Kumar, Martin R. Oswald, Erik Sandstroem, Cristian Sminchisescu, Luc van Gool: A Real-Time Learning Framework for Joint 3D Reconstruction and Semantic Segmentation. Robotics and Automation Letters Submission
3DMV, FTSDF0.501 1040.558 1050.608 1090.424 1160.478 970.690 1020.246 1120.586 960.468 940.450 910.911 1030.394 910.160 1010.438 960.212 1120.432 1080.541 1100.475 1020.742 1000.727 1000.477 104
PCNN0.498 1050.559 1040.644 1020.560 1050.420 1050.711 1010.229 1140.414 1070.436 1020.352 1060.941 780.324 1030.155 1020.238 1120.387 1040.493 1010.529 1110.509 930.813 930.751 950.504 99
3DMV0.484 1060.484 1120.538 1140.643 940.424 1040.606 1140.310 1030.574 980.433 1040.378 1010.796 1150.301 1050.214 780.537 790.208 1130.472 1060.507 1140.413 1120.693 1050.602 1140.539 91
Angela Dai, Matthias Niessner: 3DMV: Joint 3D-Multi-View Prediction for 3D Semantic Scene Segmentation. ECCV'18
PointCNN with RGBpermissive0.458 1070.577 1030.611 1070.356 1180.321 1140.715 1000.299 1070.376 1110.328 1140.319 1080.944 720.285 1070.164 990.216 1150.229 1100.484 1030.545 1090.456 1050.755 980.709 1030.475 105
Yangyan Li, Rui Bu, Mingchao Sun, Baoquan Chen: PointCNN. NeurIPS 2018
FCPNpermissive0.447 1080.679 870.604 1100.578 1040.380 1070.682 1040.291 1080.106 1180.483 910.258 1160.920 1000.258 1110.025 1150.231 1140.325 1060.480 1040.560 1070.463 1040.725 1020.666 1090.231 118
Dario Rethage, Johanna Wald, Jürgen Sturm, Nassir Navab, Federico Tombari: Fully-Convolutional Point Networks for Large-Scale Point Clouds. ECCV 2018
DGCNN_reproducecopyleft0.446 1090.474 1130.623 1050.463 1120.366 1090.651 1070.310 1030.389 1100.349 1120.330 1070.937 830.271 1090.126 1070.285 1080.224 1110.350 1140.577 1040.445 1080.625 1100.723 1010.394 110
Yue Wang, Yongbin Sun, Ziwei Liu, Sanjay E. Sarma, Michael M. Bronstein, Justin M. Solomon: Dynamic Graph CNN for Learning on Point Clouds. TOG 2019
PNET20.442 1100.548 1070.548 1130.597 1020.363 1100.628 1120.300 1050.292 1130.374 1080.307 1090.881 1100.268 1100.186 900.238 1120.204 1140.407 1100.506 1150.449 1060.667 1080.620 1130.462 108
SurfaceConvPF0.442 1100.505 1090.622 1060.380 1170.342 1120.654 1060.227 1150.397 1090.367 1090.276 1120.924 970.240 1120.198 860.359 1040.262 1080.366 1110.581 1030.435 1100.640 1090.668 1080.398 109
Hao Pan, Shilin Liu, Yang Liu, Xin Tong: Convolutional Neural Networks on 3D Surfaces Using Parallel Frames.
Tangent Convolutionspermissive0.438 1120.437 1150.646 1010.474 1110.369 1080.645 1080.353 980.258 1150.282 1170.279 1110.918 1020.298 1060.147 1060.283 1090.294 1070.487 1020.562 1060.427 1110.619 1110.633 1120.352 113
Maxim Tatarchenko, Jaesik Park, Vladlen Koltun, Qian-Yi Zhou: Tangent convolutions for dense prediction in 3d. CVPR 2018
3DWSSS0.425 1130.525 1080.647 1000.522 1060.324 1130.488 1180.077 1190.712 760.353 1100.401 990.636 1190.281 1080.176 930.340 1050.565 880.175 1180.551 1080.398 1130.370 1190.602 1140.361 112
SPLAT Netcopyleft0.393 1140.472 1140.511 1150.606 990.311 1150.656 1050.245 1130.405 1080.328 1140.197 1170.927 960.227 1140.000 1190.001 1200.249 1090.271 1170.510 1120.383 1150.593 1130.699 1050.267 116
Hang Su, Varun Jampani, Deqing Sun, Subhransu Maji, Evangelos Kalogerakis, Ming-Hsuan Yang, Jan Kautz: SPLATNet: Sparse Lattice Networks for Point Cloud Processing. CVPR 2018
ScanNet+FTSDF0.383 1150.297 1170.491 1160.432 1150.358 1110.612 1130.274 1100.116 1170.411 1060.265 1130.904 1060.229 1130.079 1120.250 1100.185 1150.320 1150.510 1120.385 1140.548 1140.597 1170.394 110
PointNet++permissive0.339 1160.584 1020.478 1170.458 1130.256 1170.360 1190.250 1110.247 1160.278 1180.261 1150.677 1180.183 1160.117 1080.212 1160.145 1170.364 1120.346 1190.232 1190.548 1140.523 1180.252 117
Charles R. Qi, Li Yi, Hao Su, Leonidas J. Guibas: pointnet++: deep hierarchical feature learning on point sets in a metric space.
GrowSP++0.323 1170.114 1190.589 1120.499 1080.147 1190.555 1150.290 1090.336 1120.290 1160.262 1140.865 1130.102 1190.000 1190.037 1180.000 1200.000 1200.462 1160.381 1160.389 1180.664 1110.473 106
SSC-UNetpermissive0.308 1180.353 1160.290 1190.278 1190.166 1180.553 1160.169 1180.286 1140.147 1190.148 1190.908 1040.182 1170.064 1130.023 1190.018 1190.354 1130.363 1170.345 1170.546 1160.685 1060.278 115
ScanNetpermissive0.306 1190.203 1180.366 1180.501 1070.311 1150.524 1170.211 1170.002 1200.342 1130.189 1180.786 1160.145 1180.102 1090.245 1110.152 1160.318 1160.348 1180.300 1180.460 1170.437 1190.182 119
Angela Dai, Angel X. Chang, Manolis Savva, Maciej Halber, Thomas Funkhouser, Matthias Nießner: ScanNet: Richly-annotated 3D Reconstructions of Indoor Scenes. CVPR'17
ERROR0.054 1200.000 1200.041 1200.172 1200.030 1200.062 1200.001 1200.035 1190.004 1200.051 1200.143 1200.019 1200.003 1180.041 1170.050 1180.003 1190.054 1200.018 1200.005 1200.264 1200.082 120